Why would AI replace proper long-term testing and evaluation? Regulators won’t allow it.
As I understand it, AI is used upfront to identify new candidates and then run massive simulations on billions of molecular interactions, binding affinities, toxicity profiles, etc. to weed out failures early, saving millions (and years) that would otherwise be wasted in wet-lab or animal work.
Only the most promising survivors then proceed to the usual process of preclinical studies, followed by full human clinical trials.
I doubt AI will replace long term testing what it will do is develop new ways to treat various ailments that previously had not been thought of and cut the time to determine if a new treatment is worth pursuing.
They're not talking about using AI to replace testing and evaluation. The idea is to use AI to more quickly discover compounds that will target certain diseases.
For example, a certain type of tumor cell may have a receptor that inhibits growth. If you can find a chemical to match or shut off that receptor, it might be a good API for a drug.
You might model PK or PD results with AI, but human testing will still be required by health ministries and the FDA.